Re: NRC Study Report

[Dave Wood <dpw@thomsoft.com>]

Marin David Condic, 561.796.8997, M/S 731-93 wrote:
> 
> >        3.  Invest $15M/year for Ada infrastructure - or drop
> >                Ada requirement entirely;
> >
>     Are you sure that's enough? I think GNAT was one of the smartest
>     decisions ever made in promoting widespread use of Ada - make a
>     good quality compiler available free to anyone who wants it. If
>     DoD would drop a significant chunk of change into developing
>     support tools (targeted to at least a PC/Windows environment) and
>     again make them available in the same way GNAT is, 

Super.  Rather than strengthen the market so that existing
independent software vendors can have a reason to stay in 
the business and even expand upon what is already available, 
let's throw some tax money at a contractor who will either 
(a) drive the ISVs out of business through unfair funded 
competition or (more likely) (b) waste a lot of tax money 
making junk software.  Have we learned nothing from the 
mammoth programs of the 80's, like ALS?  Not only were they 
a collosal waste, but they also resulted in a contractor-
oriented rather than ISV-oriented Ada community, setting 
us back a decade against C/C++ and some other languages.

It's already virtually impossible to make a profit out of
compilers (worse for, but not exclusive to, Ada), meaning
ISVs need to focus on value-added tools.  You propose to
government-fund the ISVs out of that business too?

>     (Oh. Yeah.
>     Spend a few bucks for *advertizing* that fact as well!) I think

Advertising is vastly overrated and incredibly expensive.
I think our industry has more pressing problems.

>     we'd see a dramatic increase in Ada usage in many fields. Here's
>     my initial shopping list - demonstrating that $15m is probably not
>     nearly enough:

[snip]

All of these tools already exist in one form or another for
one or more platforms.  The trick is to continue to make them
better and better.  You don't do that by throwing government
money at competing freeware - quite the opposite.  Whatever
happened to the golden goose of COTS software?  I presume
that when the government pines for Ada products to be "COTS",
they have in mind something like Delphi, not something like
"Uncle Joe's Shareware Code Generator".

>     The list could go on and on. If anyone at DoD would like me to
>     help them spend some money, I'd be glad to take the job!

Please don't.  It would be better if you could come up
with a way to (a) incentivize (not coerce) government 
contractors to use Ada products, and (b) help support
the continued development and enhancement of COTS Ada 
products from existing ISVs.  

I'll give you an example of the former.  The claim is 
that using Ada requires a bit more of up-front investment, 
with the benefit of downstream maintenance savings.  Since
most contractors don't care a fig about downstream savings,
why should they bid Ada?  They should care more that using 
Ada might make their contract more expensive up front, causing
them to lose the bid.  Solution: allow a contractor proposing 
Ada to bid a higher price compared to one bidding C, without 
putting the Ada bidder at a disadvantage in winning the
contract - essentially like a golf handicap.  Being good
capitalists, the contractor will want to bid in a way
that gives them the higher up-front cash flow, thus 
providing a kind of affirmative action for Ada.

If the government really believes that Ada is technically
superior and that the investment will be recouped 
downstream, then doesn't something like this make sense?
If the government DOESN'T believe this, then what good
is Ada at all?  Why spend even one more tax dime on Ada?

I agree with you on two things though: $15M isn't very
much money, and GNAT is a net good thing for the Ada
community.  Just keep in mind that too much of a good 
thing can be counter-productive to the long-range goal.

-- Dave Wood
-- Speaking strictly for myself